fridman@cpsc.ucalgary.ca (fridman) (01/23/91)
Greeting apollo admins: I've upgraded one of our DN4500 to SR10.3. Now I want to move the master registry to it so I can upgrade the node that is the master now. I wanted to make the 10.3 node a replica, and then swap master and replica. The problem is that the new 10.3 replica has a status of uninitialized ( doing an lrep -state from rgy_admin ). Also sometimes rgy_admin thinks the network to the new node is unreachable. Clearly I've screwed up the network some how but where can I start looking for the problem??? Thanks in advance, Robert Fridman <fridman@cpsc.ucalgary.ca>
thompson@PAN.SSEC.HONEYWELL.COM (John Thompson) (01/24/91)
> <<forwarded message>> > I've upgraded one of our DN4500 to SR10.3. Now I want to move > the master registry to it so I can upgrade the node that is > the master now. > I wanted to make the 10.3 node a replica, and then swap master and > replica. > > The problem is that the new 10.3 replica has a status of uninitialized > ( doing an lrep -state from rgy_admin ). Also sometimes rgy_admin > thinks the network to the new node is unreachable. Is the new node running an 'llbd'? If not, it won't be able to register its services (rgy) with the glbd. You can start one up w/out rebooting (as root) with "/etc/server -p /etc/ncs/llbd &" If you're running a glbd on the new node, is it started ok? Are the glbd replicas' clocks skewed? (/etc/ncs/drm_admin can tell you) If they are severely out of time-synch (> 10 minutes?) they don't pass info back and forth. They might therefore not be letting the rgyd processes talk. Is the old registry running the 10.1 rgyd? (pst will show a name of rgyd.1.2 if it's the 10.2 or greater version). I didn't think that they were incompatible in _this_ form, but it would definitely cause other problems (the new rgyd can be put onto 10.1 nodes w/ no problem. Just stop the rgyd, copy over the new prog, and re-start). You might try (as root) forcing the new node to recreate its registry : $ /etc/server -p /etc/rgyd -recreate If all else fails, try trashing out the new node's /sys/registry directory, and (as root) re-start the rgyd (as root) : $ /etc/server -p /etc/rgyd -- jt -- John Thompson Honeywell, SSEC Plymouth, MN 55441 thompson@pan.ssec.honeywell.com As ever, my opinions do not necessarily agree with Honeywell's or reality's. (Honeywell's do not necessarily agree with mine or reality's, either)